Using Frequency Attention to Make Adversarial Patch Powerful Against Person Detector
This work addresses a specific vulnerability in person detectors for security applications, representing an incremental improvement over existing adversarial patch methods.
The paper tackles the problem of adversarial patches losing effectiveness on small and medium targets in object detectors due to preprocessing shrinkage, and proposes a frequency-domain attention module (FRAN) that increases attack success rates by 4.18% and 3.89% respectively over state-of-the-art methods on YOLOv3 without harming large target performance.
Deep neural networks (DNNs) are vulnerable to adversarial attacks. In particular, object detectors may be attacked by applying a particular adversarial patch to the image. However, because the patch shrinks during preprocessing, most existing approaches that employ adversarial patches to attack object detectors would diminish the attack success rate on small and medium targets. This paper proposes a Frequency Module(FRAN), a frequency-domain attention module for guiding patch generation. This is the first study to introduce frequency domain attention to optimize the attack capabilities of adversarial patches. Our method increases the attack success rates of small and medium targets by 4.18% and 3.89%, respectively, over the state-of-the-art attack method for fooling the human detector while assaulting YOLOv3 without reducing the attack success rate of big targets.